EELE 461/561 – Digital System Design Lossy Transmission Lines

Module #7 – Lossy Lines • Lossy Lines

• Topics
- When we derived Telegrapher's Equations, we made an assumption that there was no
loss in the equivalent circuit model (i.e., R=0, G=0)
1. Skin Effect
2. Dielectric Loss - This allowed us to simplify the math and come up with the following important equations
for a Lossless T-line:

• Textbook Reading Assignments

1. 9.1-9.11
L
• What you should be able to do after this module Z0  TD  LC
C
1. Describe the physical phenomenon behind Skin Effect and Dielectric Loss
2. Use a modern CAD tool to simulate the behavior of lossy transmission lines

EELE 461/561 – Digital System Design Module #7 EELE 461/561 – Digital System Design Module #7
Page 1 Page 2

Lossy Transmission Lines Lossy Transmission Lines

• Circuit Model • Circuit Model

- If the line is lossy, we need to include the series resistance and shunt conductance back into - Let's derive the relationship between Voltage & Current to time and space with the full model
our equivalent T-line circuit model
- We first define the length of the wire using dx.

- Since our electrical components are defined in unit length, the total values can be found by
multiplying the unit length value by dx.

R' = Resistance per Unit Length

L' = Inductance per Unit Length
C' = Capacitance per Unit Length
G' = Conductance per Unit Length

Lseg = L'·dx
Rseg = R'·dx
Cseg = C'·dx
Gseg = G'·dx

dx

EELE 461/561 – Digital System Design Module #7 EELE 461/561 – Digital System Design Module #7
Page 3 Page 4

Lossy Transmission Lines Lossy Transmission Lines

• Circuit Model • Circuit Model

- We enter the segment at (x) and we exit the line at (x+dx) - The input voltage can be described as: V(x,t)
- The input current can be described as: I(x,t)
- The output voltage can be described as: V(x+dx,t)
- The output current can be described as: I(x+dx,t)

I(x, t) I(x+dx, t)

+ +
V(x, t) V(x+dx, t)
- -

dx
dx

EELE 461/561 – Digital System Design Module #7 EELE 461/561 – Digital System Design Module #7
Page 5 Page 6

1
 Lossy Transmission Lines Lossy Transmission Lines

• Circuit Model • Circuit Model

- We can write an expression for the voltage drop across the inductor and resistor using KVL: - Now we can write an expression for the output current using KCL:

V ( x, t )  V ( x  dx, t )  R' dx  I ( x, t )  L' dx 

dI ( x, t ) dV ( x  dx, t )
I ( x, t )  I ( x  dx, t )  G' dx V ( x  dx, t )  C ' dx 
dt dt
- Which is rewritten as:
- Which can be rewritten as:

V ( x  dx, t )  V ( x, t )  dI ( x, t )  I ( x  dx, t )  I ( x, t )  dV ( x  dx, t )

   R'I ( x, t )  L' dt 
 dx  dx   G'V ( x  dx, t )  C ' dt

- If we let dx→0, we are left with a differential equation: - If we let dx→0, we are left with another differential equation:

dV ( x, t ) dI ( x, t ) dI ( x, t ) dV ( x, t )
  R'I ( x, t )  L'   G'V ( x, t )  C '
dx dt dx dt

EELE 461/561 – Digital System Design Module #7 EELE 461/561 – Digital System Design Module #7
Page 7 Page 8

Lossy Transmission Lines Lossy Transmission Lines

• Circuit Model • Wave Equations

- These two 1st order differential equations describe the complete interaction of V and I on a T-line. - We can put the voltage expression into the form of the Wave Equation by differentiating the first
(i.e., "Telegrapher's Equations") Telegrapher equation with respect to x

differentiate with respect to x

dV dI dI dV
  R'I  L'   G'V  C '
dx dt
 R' jL'  I ( x)
dx dt dV ( x)

dx

- Let's put these into a more usable form using phasor representation where (d/dt → j)
d 2V ( x)
 R' jL' 
dI ( x)

dx 2 dx
 R' jL'  I ( x)  G' jC ' V ( x)
dV ( x) dI ( x)
 
dx dx

EELE 461/561 – Digital System Design Module #7 EELE 461/561 – Digital System Design Module #7
Page 9 Page 10

Lossy Transmission Lines Lossy Transmission Lines

• Wave Equations • Wave Equation

- We can now plug in our expression for the derivative of current: - If we define g as:

2  R' jL' G' jC'

 G' jC ' V ( x)
dI ( x)
 R' jL' 
d V ( x) dI ( x)
 
dx 2 dx dx
- we can rewrite the voltage expression in Wave Equation form as:

d 2V ( x)
  R' jL'  G ' jC '  V ( x)
dx 2
d 2V ( x)
  2  V ( x)  0
- Rearranging, we get: dx 2

d 2V ( x)
 R' jL'  G ' jC ' V ( x)  0
dx 2

EELE 461/561 – Digital System Design Module #7 EELE 461/561 – Digital System Design Module #7
Page 11 Page 12

2
 Lossy Transmission Lines Lossy Transmission Lines

• Wave Equations • Wave Equations

- Now let's put the current expression into the form of the Wave Equation by differentiating the second - We can now plug in our expression for the derivative of voltage:
Telegrapher equation with respect to x

d 2 I ( x)
 G ' jC ' 
dV ( x)
 R' jL'  I ( x)
differentiate with respect to x
 dV ( x)

dx 2 dx dx
 G ' jC '  V ( x)
dI ( x)

dx
d 2 I ( x)
  G ' jC '  ( R' jL' )  I ( x)
dx 2

d 2 I ( x) - Rearranging, we get:
 G ' jC ' 
dV ( x)

dx 2 dx
d 2 I ( x)
 R' jL'  G ' jC '  I ( x)  0
dx 2

EELE 461/561 – Digital System Design Module #7 EELE 461/561 – Digital System Design Module #7
Page 13 Page 14

Lossy Transmission Lines Lossy Transmission Lines

• Wave Equations • Wave Equations

- Since, we have already defined g as: - These two 2nd order differential equations describe the voltage and current propagation on a
lossy transmission line:
 R' jL' G' jC'

d 2V ( x) d 2 I ( x)
- we can rewrite the current expression in Wave Equation form as:   2  V ( x)  0   2  I ( x)  0
dx 2 dx 2

d 2 I ( x) - We call g the "Complex Propagation Constant"

  2  I ( x)  0
dx 2
 R' jL' G' jC'
which consists of a Real and Imaginary part:

    j

EELE 461/561 – Digital System Design Module #7 EELE 461/561 – Digital System Design Module #7
Page 15 Page 16

Lossy Transmission Lines Lossy Transmission Lines

• Attenuation Constant • Phase Constant

- We call the Real part of g the Attenuation Constant () - We call the Imaginary part of g the Phase Constant ( )

  Re( )  Re  R' jL' G' jC'

  Im( )  Im  R' jL' G' jC'
- This quantity has units of rad/m

- This can be used to calculate the wave velocity with:

- This quantity has units of Np/m 
velocity   p 
NOTE: A Napier is unit of ratio. It is used to describe ratios of voltages and currents. 
• Complex Propagation Constant for Passive T-lines
x 
Np  ln  1  - We choose the square root values of g that give positive values for  and .
 x2 
    j
- Remember that in a Lossless line, =0.
- For passive T-lines,
- In a Lossy line 0
 0

EELE 461/561 – Digital System Design Module #7 EELE 461/561 – Digital System Design Module #7
Page 17 Page 18

3
 Lossy Transmission Lines Lossy Transmission Lines

• Characteristic Impedance • Characteristic Impedance

- The Wave Equations have traveling wave solutions for a lossy medium in the form of: - If we differentiate the voltage solution, we can get the solution for current by plugging it back
into our original Telegrapher's Equation:
 x   x  
V ( x)  V e V e  V ( x)  V ( x)
I ( x)  I  e x  I  e x  I  ( x)  I  ( x) V ( x)  V  e x  V  e x

dV
 V  e x  V  e x  ( R' jL' )  I ( x)
Forward Reverse dx
Traveling Traveling
Wave Wave
dV
dx
 
  V  e x  V  e x  ( R' jL' )  I ( x)

- Rearranging for I, we get:


I ( x) 
R' jL'

 V  e x  V  e x 

EELE 461/561 – Digital System Design Module #7 EELE 461/561 – Digital System Design Module #7
Page 19 Page 20

Lossy Transmission Lines Lossy Transmission Lines

• Characteristic Impedance • Characteristic Impedance

- Remembering the definition for impedance is the ratio of either the forward or reverse traveling waves: - Now we have our expression of the characteristic impedance of a lossy transmission line:

V  V 
Z  
I I R' jL'
Z0 
- We can now plug in our equations for V(x) and I(x), and only consider the forwarding traveling waves: G ' jC '

Z0 
V

V   x
e  V  e x  
V   x
e 

  
 
- Notice that this quantity is complex. This implies that the magnitude of the impedance depends
I  V  e x  V  e x  V  e x on the incidence waves frequency.
R' jL' R' jL'
- When a digital signal is traveling in a lossy line, each harmonic component of the square wave
will experience a different impedance.
R' jL' R' jL' R' jL'  R' jL'
Z0   
 R' jL'  G ' jC ' R' jL'  G ' jC '
R' jL' R' jL'
Z0  
G ' jC ' G ' jC '

EELE 461/561 – Digital System Design Module #7 EELE 461/561 – Digital System Design Module #7
Page 21 Page 22

Sources of Loss Conductor Loss

• Sources of Loss • Skin Depth

- Loss refers to the amount of signal transmitted that does not reach the receiver. - At DC, the resistance of a conductor is proportional to:

- Loss can occur due to many sources: - the cross-sectional area of the conductor
- the bulk resistivity of the material
1) Impedance Mismatches Leading to Reflected Energy
2) Coupling to adjacent Traces
L L
3) Radiation Loss R 
4) Conductor Loss W  H Area
5) Dielectric Loss

- When we talk about Lossy Transmission Lines, we tend to focus on items 4 & 5

- Reflections & Coupling can be modeled using our standard LC model of a T-line. - At DC, the charge is equally distributed across the cross-section of the conductor:

- Radiation is typically small, and is difficult to model using circuit models so we treat
that using EM simulators.

- Conductor Loss and Dielectric Loss can be modeled using our complete RLCG transmission line
circuit.

- When people talk about Lossy Lines, they are typically referring to Conductor & Dielectric Loss

EELE 461/561 – Digital System Design Module #7 EELE 461/561 – Digital System Design Module #7
Page 23 Page 24

4
 Conductor Loss Conductor Loss

• Skin Depth • Skin Depth

- When AC current flows through the conductor, the charge is not equally distributed within the - The phenomenon of the current flowing through this reduced cross-section of the
cross-sectional area of the conductor. At AC, the current will attempt to find the path conductor is described using a quantity called Skin Depth
of least impedance
- The reduced cross-section has the effect of increasing the series resistance of the conductor
- This results in two trends: as the frequency increases.

1) The current within the conductor will spread out as far as possible in order to minimize
the partial self-inductance of the conductor:

2) The current within the conductor will try to move as close as possible to the oppositely
directed return current in more to maximum the partial mutual inductance between the
two currents.

(1) (2)

EELE 461/561 – Digital System Design Module #7 EELE 461/561 – Digital System Design Module #7
Page 25 Page 26

Conductor Loss Conductor Loss

• Skin Depth • Skin Depth

- Skin Depth is described by  and is expressed as: - Skin Depth directly effects the series resistance of a Transmission line by reducing the
effective cross-sectional area of the conductor that the current can flow through.

2   L
  
1 RS 
    f Area ( )
    f 
- Notice that the skin depth is inversely proportional to the square root of frequency:
where:  is the resistivity of the conductor
 is the conductivity of the conductor (1/ ) 1 1
 is the magnetic permeability (0·r)  
    f  f
- Skin depth has units of meters and is the definition of the depth below the surface of a conductor
with depth (d) at which the current density (J) decays to 1/e (~0.37) of the current
density of the surface (JS): d / - Since skin depth is in the denominator of the expression for series resistance of the conductor,
J  JSe this means that the resistance is proportional to the square root of frequency:

RS  f

EELE 461/561 – Digital System Design Module #7 EELE 461/561 – Digital System Design Module #7
Page 27 Page 28

Conductor Loss Conductor Loss

• Skin Depth • Total Conductor Loss

- We model the conductor loss due to - The complete expression for conductor loss should also include any DC loss due to the
skin depth using the series resistor resistivity of the bulk material:
in our RLGC T-line model.

- This resistance value is dependant on

frequency:
Rconductor  RDC  RAC f
RS ( f )  f

- Note that it is hard to define one complete expression for the AC conductor resistance because
- When stimulating the lossy transmission line it depends on the geometry of the conductor. (i.e., round, square, rectangle, etc…)
with a digital signal, each frequency component
of the signal will experience a different - For a given cross sectional shape, the skin depth is then applied to that shape in order to predict
series resistance. the new cross-sectional area that the current flows through.

EELE 461/561 – Digital System Design Module #7 EELE 461/561 – Digital System Design Module #7
Page 29 Page 30

5
 Lossy Transmission Lines Lossy Transmission Lines

• Dielectric Loss • Dielectric Loss

- Now we move to the 2nd main effect in lossy lines, the dielectric. - An ideal capacitor will store a particular amount of charge depending on the voltage applied.

- We use the G' element in our model to account for dielectric loss.
Q
C
V

- The capacitor is constructed using two conducting surfaces separated by a dielectric.

- In the ideal case, this structure does NOT allow DC current to flow between the two terminals of
the device.

EELE 461/561 – Digital System Design Module #7 EELE 461/561 – Digital System Design Module #7
Page 31 Page 32

Lossy Transmission Lines Lossy Transmission Lines

• Dielectric Loss • Dielectric Loss

- In reality, the charge in the capacitor is held by electric dipoles within the material. - Since there is no DC current (or no in-phase current), there is no real power dissipation in
an ideal capacitor.
- If an AC voltage is applied to the capacitor, the dipoles will align to the direction of the
applied electric field. - This allows us to say that the resistivity of an ideal capacitor is .

- This movement of charge results in AC current flow. - However, real capacitors do have some resistance associated with them.

- This means that current will flow through the capacitor that is in phase with the voltage.

- We model this resistance with a resistor component in parallel with the capacitor.

- The amount of current that flows is proportional to the rate-of-change of voltage across the capacitor:

dV
ic  C
dt

EELE 461/561 – Digital System Design Module #7 EELE 461/561 – Digital System Design Module #7
Page 33 Page 34

Lossy Transmission Lines Lossy Transmission Lines

• Dielectric Loss • Dielectric Loss

- The density of available dipoles in a material to hold the charge is reflected in the dielectric - The dipoles in the material do not re-orient instantaneously upon a change in voltage.
constant (i.e., the electric permittivity).
- For a given time varying voltage, there are going to be dipoles that are perfectly aligned with
- The density of dipoles available in air is described with a permittivity of 0 the applied electric field. These dipoles contribute to the capacitance of the structure and
result in an out of phase current (-90)
- If a parallel plate capacitor was constructed using an air dielectric, the capacitance would be
given by: - Due to the finite speed at which the dipoles can change, there will be dipoles that are aligned
0  A perpendicular to the applied electric field which produce an in phase current
C0 
t - This in phase current is the source of the leakage current in a capacitor

- If the air was then replaced with a different insulating material with r>1, the new capacitance
would be described as:

C   r  C0

EELE 461/561 – Digital System Design Module #7 EELE 461/561 – Digital System Design Module #7
Page 35 Page 36

6
 Lossy Transmission Lines Lossy Transmission Lines

• Dielectric Loss • The Complex Dielectric Constant

- If we apply an arbitrary sinusoidal voltage to a capacitor: - We now can use the relative permittivity of the material to describe both the in phase current and
the out of phase current by making it a complex quantity:
V  V0  e jt
 r   r ' j r ' '
the current that results is described as: where:

r = the complex dielectric constant

dV
I C  jCV
dt r' = the real part of the complex dielectric constant, representing the
out of phase current flow in the capacitor. This is what we have always
called simply the dielectric constant. Remember that this current is always
- We can use our definition for a non-air dielectric capacitor to include the dielectric constant defined as being -90o out of phase with the voltage
in this equation:
r'' = the imaginary part of the complex dielectric constant, representing the
in phase current flow in the capacitor. This is the new quantity that we've added
dV
I C  j r C0V to represent the current due to the loss in the dielectric. Since this current
dt is in phase, we need to apply a "-" sign in the complex quantity to re-align it
back to the voltage (remember by definition the current in the capacitor starts
at -90o)

EELE 461/561 – Digital System Design Module #7 EELE 461/561 – Digital System Design Module #7
Page 37 Page 38

Lossy Transmission Lines Lossy Transmission Lines

• The loss angle () • The Loss Tangent, (tan())

- we describe the angle between the real and imaginary parts of the complex dielectric constant - By convention, we use the ratio of the imaginary to the real parts of the dielectric constant to
as the loss angle () describe the loss in a material:

- Note: this is NOT the same as the skin depth, it is just an unfortunate coincidence that delta is - This is called the Loss Tangent () or Dissipation Factor
used for both. Sometimes people will use  for the loss angle

Im Im
| r | r '' | r |
tan( ) 
r '
 r   r ' j r ' ' r '' r ''
 
Re Re
r '  r ' '   r ' tan( ) r '

EELE 461/561 – Digital System Design Module #7 EELE 461/561 – Digital System Design Module #7
Page 39 Page 40

Lossy Transmission Lines Lossy Transmission Lines

• The Loss Tangent (Dissipation Factor) • The Loss Tangent (Dissipation Factor)

- At this point, things can get a little confusing. Remember that: - Now let's describe the loss in the dielectric using the Loss Tangent.

- The complete expression for the current in our capacitor is given by:

 
- the dielectric constant is described as a complex quantity in order to describe both:
I  j r C0V  j  r  j r C0V  j r C0V   r C0V   r C0V  j r C0V
' '' ' '' '' '

- the out of phase current due to an applied voltage

I (Re)   r C0V
''

- the in phase current due to an applied voltage

I (Im)   r C0V
'

- The out of phase current that results from an applied voltage is actually the expected
response of an ideal capacitor. That is why this quantity is described using the
real part of the complex dielectric constant. - The resistance of the material is described as:

- The in phase current that results from an applied voltage is due to the loss in a
real capacitor and is described using the imaginary part of the complex dielectric constant. V V 1 1 1
Rleakage     
I (Re)  r ''C0V  r ''C0  r ' tan( )C0  tan( )C

EELE 461/561 – Digital System Design Module #7 EELE 461/561 – Digital System Design Module #7
Page 41 Page 42

7
 Lossy Transmission Lines Lossy Transmission Lines

• The Loss Tangent (Dissipation Factor) • The Loss Tangent (Dissipation Factor)

- We can now put his expression for Rleakage into terms of bulk conductivity of the material: - Now we have the bulk conductivity (and resistivity) in terms of the Loss Tangent and frequency:

1    0 r '  tan( )
Rleakage 
 tan( )C 
1
 0 r '  tan( )
- The definition of resistance & capacitance of a test structure is (looking from the top to bottom):
- This can we used to describe the conductance or resistance of the shunt resistor in our model:
  Length 1 H  0   r '  Area Area
R   C  0 r 
'

Area  Area H H G'  f

- These two expressions can be combined to form a relationship between the loss tangent & the 1
bulk conductivity of the material: R' 
f
H 11  ' 1
R    0 r 
 Area  C  tan( )C
- We typically use G' to model the dielectric loss because it scales directly with length. That way
all 4 of our T-line parameters will scale with length
   0 r '  tan( )

EELE 461/561 – Digital System Design Module #7 EELE 461/561 – Digital System Design Module #7
Page 43 Page 44

Lossy Transmission Lines Lossy Transmission Lines

• Dielectric Loss • Total Loss

- We model the conductor loss due to - we now see that there are two frequency dependant sources of loss in our lossy T-line.
skin depth using the shunt conductance
in our RLGC T-line model. - This results in higher frequencies being attenuated more than lower frequencies.

- This conductance value is dependant on - The two frequency dependant sources are spec'd in terms of unit length (i.e., R', G')
frequency:
Gdielectric  2  f  tan( )C
Gshunt( f )  f
RS  RDC  RAC f

- When stimulating the lossy transmission line

with a digital signal, each frequency component RS ( f )  f
of the signal will experience a different
shunt conductance.

Gshunt( f )  f

EELE 461/561 – Digital System Design Module #7 EELE 461/561 – Digital System Design Module #7
Page 45 Page 46

Lossy Transmission Lines Lossy Transmission Lines

• Signal Velocity & Dispersion • Approximations

- we have seen that the velocity of a sine wave in a lossy medium depends on the frequency. - There are a couple approximations that can be made to simplify the analysis of lossy lines:

 Low Loss Regime

up 
 - when RS<<XL and GL<<XC

- we can use the lossless equations for Z0 and velocity

- For a digital signal, this means that different frequency components will travel through the lossy
line at different rates. Good Conductor

- The end result will be a distorted signal at the receiver. - when RS<<XL (i.e., skin effect is negligible)

- This phenomenon is called Dispersion

  f   f
Low Loss Dielectric

- when GL<<XC (i.e, the dielectric leakage is negligible)

     

2 
EELE 461/561 – Digital System Design Module #7 EELE 461/561 – Digital System Design Module #7
Page 47 Page 48